Machine learning models based on artificial intelligence-derived CT quantitative parameters and CT semantic features for predicting the invasive risk of subsolid pulmonary nodules: a dual-center study
Abstract
PurposeTo develop and validate machine learning (ML) models integrating artificial intelligence (AI)-derived quantitative CT parameters and CT semantic features for predicting the invasiveness risk of subsolid pulmonary nodules (SSNs).Materials and methodsThis retrospective dual-center study enrolled 833 patients with surgically resected, pathologically confirmed adenocarcinoma-spectrum SSNs (training, n=586; external validation, n=247). Low risk was defined as atypical adenomatous hyperplasia (AAH), adenocarcinoma in situ (AIS), or minimally invasive adenocarcinoma (MIA); high risk as invasive adenocarcinoma (IAC). The least absolute shrinkage and selection operator (LASSO) selected features from 19 AI parameters, 7 semantic features, and 3 clinical characteristics. Seven ML models were built and validated using ten-fold cross-validation and external validation. SHapley Additive exPlanations (SHAP) provided interpretability.ResultsLASSO identified nine predictors (margin, spiculation, lobulation, pleural indentation, vascular convergence, long diameter, solid component ratio, skewness, entropy). In internal cross-validation, all seven models showed comparable performance (mean AUC 0.935–0.940). In external validation, random forest (RF) showed the numerically highest AUC (0.959, 95% CI: 0.938–0.980), although pairwise DeLong tests with Bonferroni correction showed no statistically significant differences among models. SHAP analysis identified long diameter, skewness, spiculation, entropy, and solid component ratio as the top five predictors contributing to invasive risk. An online tool based on RF was deployed at https://subsolid-nodule.shinyapps.io/data/ for real-time individual risk calculation.ConclusionEnsemble learning models, particularly RF, integrating AI-derived CT quantitative parameters and CT semantic features accurately predict SSN invasive risk. This non-invasive tool may provide additional information for preoperative risk stratification and multidisciplinary assessment.
Related
- Diagnostic performance of spectral CT quantitative parameters in predicting spread through air spaces of lung adenocarcinoma
- Systematic review of spectral computed tomography quantitative parameters for predicting EGFR mutation status in non-small cell lung cancer
- Quantitative CT-derived small pulmonary vessel and erector spinae muscle parameters associated with acute exacerbation in COPD patients from a cold region: a dual-center study
- Predicting follow-up cognition using estimated structural and functional connectivity networks and artificial intelligence in multiple sclerosis
- Artificial intelligence-assisted bronchoscopy for the diagnosis and treatment of pulmonary diseases: Advances and clinical practice
BibTeX
@article{she2026machine,
title = {{Machine learning models based on artificial intelligence-derived CT quantitative parameters and CT semantic features for predicting the invasive risk of subsolid pulmonary nodules: a dual-center study}},
author = {Lilan She and Tengyong Yang and Jiaxin Wu and Guiling Chen and Meilan Huang and Xuhui Chen and Yunjing Xue},
journal = {Frontiers in Oncology},
year = {2026},
month = nov,
doi = {10.3389/fonc.2026.1950196},
url = {https://www.frontiersin.org/articles/10.3389/fonc.2026.1950196},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Machine learning models based on artificial intelligence-derived CT quantitative parameters and CT semantic features for predicting the invasive risk of subsolid pulmonary nodules: a dual-center study
AU - Lilan She
AU - Tengyong Yang
AU - Jiaxin Wu
AU - Guiling Chen
AU - Meilan Huang
AU - Xuhui Chen
AU - Yunjing Xue
PY - 2026
DA - 2026/11/03
JO - Frontiers in Oncology
DO - 10.3389/fonc.2026.1950196
UR - https://www.frontiersin.org/articles/10.3389/fonc.2026.1950196
AB - PurposeTo develop and validate machine learning (ML) models integrating artificial intelligence (AI)-derived quantitative CT parameters and CT semantic features for predicting the invasiveness risk of subsolid pulmonary nodules (SSNs).Materials and methodsThis retrospective dual-center study enrolled 833 patients with surgically resected, pathologically confirmed adenocarcinoma-spectrum SSNs (training, n=586; external validation, n=247). Low risk was defined as atypical adenomatous hyperplasia (AAH), adenocarcinoma in situ (AIS), or minimally invasive adenocarcinoma (MIA); high risk as invasive adenocarcinoma (IAC). The least absolute shrinkage and selection operator (LASSO) selected features from 19 AI parameters, 7 semantic features, and 3 clinical characteristics. Seven ML models were built and validated using ten-fold cross-validation and external validation. SHapley Additive exPlanations (SHAP) provided interpretability.ResultsLASSO identified nine predictors (margin, spiculation, lobulation, pleural indentation, vascular convergence, long diameter, solid component ratio, skewness, entropy). In internal cross-validation, all seven models showed comparable performance (mean AUC 0.935–0.940). In external validation, random forest (RF) showed the numerically highest AUC (0.959, 95% CI: 0.938–0.980), although pairwise DeLong tests with Bonferroni correction showed no statistically significant differences among models. SHAP analysis identified long diameter, skewness, spiculation, entropy, and solid component ratio as the top five predictors contributing to invasive risk. An online tool based on RF was deployed at https://subsolid-nodule.shinyapps.io/data/ for real-time individual risk calculation.ConclusionEnsemble learning models, particularly RF, integrating AI-derived CT quantitative parameters and CT semantic features accurately predict SSN invasive risk. This non-invasive tool may provide additional information for preoperative risk stratification and multidisciplinary assessment.
SN - 2234-943X
ER -